[PATCH v4 5/6] xfs: add quota-driven speculative preallocation throttling

[Mark Tinguely]

On 02/25/13 16:14, Dave Chinner wrote:
> On Mon, Feb 25, 2013 at 03:44:10PM -0600, Mark Tinguely wrote:
>> On 02/20/13 09:10, Brian Foster wrote:
>>> Introduce the need_throttle() and calc_throttle() functions to
>>> independently check whether throttling is required for a particular
>>> dquot and if so, calculate the associated throttling metrics based
>>> on the state of the quota. We use the same general algorithm to
>>> calculate the throttle shift as for global free space with the
>>> exception of using three stages rather than five.
>>>
>>> Update xfs_iomap_prealloc_size() to use the smallest available
>>> prealloc size based on each of the constraints and apply the
>>> maximum shift to obtain the throttled preallocation size.
>>>
>>> Signed-off-by: Brian Foster<bfoster at redhat.com>
>>> ---
>>
>>>   	/*
>>>   	 * MAXEXTLEN is not a power of two value but we round the prealloc down
>>> @@ -412,6 +472,28 @@ xfs_iomap_prealloc_size(
>>>   		if (freesp<   mp->m_low_space[XFS_LOWSP_1_PCNT])
>>>   			shift++;
>>>   	}
>>> +
>>> +	/*
>>> +	 * Check each quota to cap the prealloc size and provide a shift
>>> +	 * value to throttle with.
>>> +	 */
>>> +	if (xfs_quota_need_throttle(ip, XFS_DQ_USER, alloc_blocks))
>>> +		xfs_quota_calc_throttle(ip, XFS_DQ_USER,&qblocks,&qshift);
>>> +	if (xfs_quota_need_throttle(ip, XFS_DQ_GROUP, alloc_blocks))
>>> +		xfs_quota_calc_throttle(ip, XFS_DQ_GROUP,&qblocks,&qshift);
>>> +	if (xfs_quota_need_throttle(ip, XFS_DQ_PROJ, alloc_blocks))
>>> +		xfs_quota_calc_throttle(ip, XFS_DQ_PROJ,&qblocks,&qshift);
>>> +
>>> +	/*
>>> +	 * The final prealloc size is set to the minimum of free space available
>>> +	 * in each of the quotas and the overall filesystem.
>>> +	 *
>>> +	 * The shift throttle value is set to the maximum value as determined by
>>> +	 * the global low free space values and per-quota low free space values.
>>> +	 */
>>> +	alloc_blocks = MIN(alloc_blocks, qblocks);
>>> +	shift = MAX(shift, qshift);
>>> +
>>>   	if (shift)
>>>   		alloc_blocks>>= shift;
>>>   	/*
>>
>> All the limits are applied from previous extents, quota and then the
>> code from commit 055388a3 is applied:
>> 	if (alloc_blocks<  mp->m_writeio_blocks)
>> 		alloc_blocks = mp->m_writeio_blocks;
>>
>>                         ^^^^
>> we may not have mp->m_writeio_blocks left in the filesytem.
>> Doesn't the following make more sense?:
>>
>> 	if (alloc_blocks<  mp->m_writeio_blocks)
>> 		alloc_blocks = 0;
>
> No. This is for prealloc, and we alwys try to prealloc the minimum
> configured. If we can't prealloc that amount, then preallocation
> fails and we'll try a single block.
>
> Remember, the freesp calculation is done unlocked and hence freesp
> is only an estimate. It may change between the reading of it and the
> actual allocation attempt, and so the only real determination of
> ENOSPC is the failure ot the allocation attempt.....
>
> Cheers,
>
> Dave.

Thank-you for the information.

And without the minimum, it will immediately trip the assert:

Assertion failed: last_fsb > offset_fsb, file: 
/root/xfs/fs/xfs/xfs_iomap.c, line: 590

Sorry for the noise.

--Mark.